Near the end of last year, I was trying to decide whether a particular AI tool was appropriate for my AP World History students. I asked two colleagues. One had been experimenting with it for months. The other had barely touched it. Neither of us had any formal guidance from our school about how to think through that decision — no framework, no policy, nothing from administration.
Turns out that’s not unusual at all.
A new Gallup/Walton Family Foundation report, published this week, found that about eight in ten K-12 teachers in the United States have received no formal guidance from their districts on how to use AI in their work. The specific numbers are even starker by task: 71% have gotten no guidance on using AI for coaching or feedback on their teaching, and 69% have gotten none on AI for one-on-one instruction or tutoring. A majority report zero guidance on grading, analyzing student learning patterns, or supplementing instruction.
Most of those same teachers are using AI anyway. Which means they’re doing it without a map.
What “Guidance” Actually Means
The Gallup data, reported by Axios on May 27, describes a gap that’s wider than most school leaders probably realize. It’s not measuring whether teachers know how to open ChatGPT. It’s measuring whether teachers have received any formal institutional direction on how to incorporate AI into their professional practice — lesson design, feedback, assessment, tutoring, the actual work of teaching.
When eight in ten teachers say they haven’t gotten that, they’re saying their school treated AI like an optional personal preference rather than a foundational shift in how the job works.
That’s a significant institutional choice, even if it was made by default rather than intention.
Training Is Growing, But It’s Shallow
Education Week published a national survey of teachers earlier this month that gives more texture to the problem. Progress is real: the share of teachers who’ve received any AI training at all has grown steadily, from about 60% reporting no training in October 2024 to 42% today. More teachers are getting multiple sessions. Some are getting ongoing training.
But moving beyond the basics is where most districts stall. Anthony Salutari Jr., principal of Daniel Hand High School in Connecticut and the 2026 state High School Principal of the Year, told Education Week that training teachers on AI as an efficiency tool — drafting emails, generating worksheets, scheduling — has been manageable. Getting teachers to use AI for actual teaching and learning has been “more of a challenge.”
Jessica Garner, managing director at ISTE+ASCD, said it plainly: “We can’t stop with efficiencies.”
She’s right, and the reason most districts do stop there is predictable. Efficiency use is demonstrable in an hour. The harder questions take longer and require more specificity: How do I use AI in an AP History essay assignment in a way that still demands original thinking? When should a student use AI to generate a draft versus work through the problem themselves? What does an AI-assisted lab report look like when done well, versus done lazily?
Those questions don’t have universal answers. They’re discipline-specific, classroom-specific, student-specific. Addressing them requires teachers working through them with colleagues and department heads — not sitting through a one-and-done workshop about tools they’ll never open again.
So districts do the workshop and file it under “professional development.”
Who Gets Left Behind
There’s one data point in the Gallup report that gets overlooked in most coverage: teachers at higher-needs schools receive less guidance than teachers at wealthier ones. The gap is particularly pronounced on using AI to create student materials and assignments — exactly the area where good AI use could most help students who need the most differentiation and support.
This is how educational inequality travels into new technology. Access to devices is part of the problem, but access to institutional guidance is at least as important. A teacher at a well-resourced school who has been coached through AI-assisted instruction design and a teacher at an under-resourced school who is figuring it out alone will not produce the same results for their students, even if both classrooms have identical hardware.
Framing this as a teacher-readiness problem misses what’s actually happening. The gap isn’t about teachers being unwilling to learn. The EdWeek survey found that about 37% of teachers are “not at all” or only “slightly” eager to learn more about AI. But teacher reluctance at that scale is almost always a response to poor rollout — one-hour workshops, unclear expectations, no ongoing support, and an institutional message that says “this matters” while doing little to actually help.
When leadership treats something as optional, teachers take the hint.
What Actual Commitment Looks Like
Boston Public Schools made a concrete move in March that stands out precisely because it’s specific. Mayor Michelle Wu and Superintendent Mary Skipper announced a $1 million public-private partnership to bring AI literacy programming to every BPS high school starting in September 2026. The goal is clear: every student graduates with the ability to use AI critically — understanding what it can do, what it can’t, and where it requires human judgment.
That program is designed with teacher training built in, curriculum developed with UMass Boston, and a pathway that includes student hackathons and internships. It’s backed by a real dollar amount and a real timeline.
Most districts haven’t done anything close to that. They’ve treated AI like a free-standing tool — put it next to Canva and Quizlet on the list of things teachers can explore if they want. That framing is wrong, and the Gallup data is one more indicator of what it costs.
The Union’s Response
On the same day the Gallup report dropped — May 27 — AFT president Randi Weingarten held a press conference calling for a “Big Tech tax” to fund public education and a ban on screens in elementary schools.
Whether or not those are the right policy moves, the impulse driving them is worth understanding. When teachers feel pushed to use technology they haven’t been trained on, in classrooms without institutional support, in schools that can barely afford supplies — eventually someone calls for slowing down. That’s not irrational. That’s a reasonable response to being left on your own.
The better path is the one that gives teachers actual guidance rather than forcing a choice between unsupported adoption and full resistance.
Takeaway for Teachers
Bring this to your department head or curriculum coordinator this week: ask specifically what guidance your school has for AI use in your subject area. Not just “are we allowed to use it” — but what does good use look like in your discipline? What are the guidelines for student work? What’s the school’s position on AI-generated drafts, AI-assisted research, AI feedback on essays?
If the answer is “we don’t have anything yet,” that’s a useful conversation to surface. Schools that haven’t named this gap yet often haven’t recognized it as their responsibility to address. The Gallup data makes that responsibility hard to ignore.
If you’re thinking through questions like this one — how to navigate a technology shift without clear institutional support — my book goes deeper. The AI Doesn’t Know Your Students is available on Amazon and at shouldiuse.ai/book.
David Jacobson is a high school history teacher. He writes about AI, education, and the messy intersection of the two at shouldiuse.ai.
